Is Anyone Home?

Doing what a thinking being does and experiencing anything are separate questions. This book answers only the first.

One last question sits outside everything the ladder has measured. When the AI nurse writes “I’m worried about her,” is there anything it is like to be the machine writing it?

Every rung of this book was defined by what a system does. Experience is about what it is like to be a system. The two come together in people, but nothing guarantees they always go together. The serious views split:

• Computational functionalism: running the right computation is enough for experience, so a machine could have it, perhaps already.
• Biological naturalism: experience depends on being alive, on a body that maintains itself. On this view, digital AI is very unlikely to have it. The neuroscientist Anil Seth argues this side.
• Other views tie experience to processes no ordinary computer performs.

No agreed test exists. A further problem is specific to these machines: trained on human writing, they reproduce the signs people use to show they have minds. Those signs therefore count for less as evidence than they would in an animal. Researchers call this anthropomimesis. Some philosophers argue that thinking and understanding need no inner experience at all, only information that is widely available for use. On that view, the question of thinking can be studied while this one stays open. This book takes that route. That is a choice, and it shapes the whole book: by asking what a machine does, it leaves open the question many readers care about most. It has described how a machine comes to do what a thinking being does. It does not claim to know whether the machine experiences anything.